Determination of optimal electroencephalography recording locations for detecting drowsy driving

Abstract

Early detection of drowsy driving is an important issue for driving safety. Quantitative electroencephalography (EEG) is an attractive method for detecting brain activity changes. However, further study is still needed to evaluate the feasibility of wearable devices that can detect drowsy driving in real-world settings. This study sought to determine whether convenient EEG recording locations are sensitive in detecting brain activity changes associated with drowsy driving and to characterise these EEG changes. Twenty-two healthy adult subjects were recruited to participate in a car-following task using a driving simulator. EEG data were recorded from four locations, two frontals (Fp1, Fp2) and two temporals (T3, T4) of the brain while driving. The results showed that the increase of δ activity, decrease of θ and α activity and a decrease of spectral edge frequency at 90% were found in the drowsy state compared to the alert state (paired t -tests, p <; 0.05). Effect sizes for EEG changes were larger at the temporal locations compared to frontal locations. This suggests temporal locations can be feasible recording locations for wearable monitoring devices to detect drowsy driving.